State of Software Development - Research Paper Example

Of Software Development By of of of Department] This research analyzes the relationship between prior research findings and current analytical and statistical findings of factors regarding software developers and the factors affecting them. The factors under consideration in this research are the effects of job satisfaction, well being and agility upon software developers. The analytical research involved collection of data from ninety eight software developers. The statistical analysis resulted in the findings that there were no significant relationship among agility and the job satisfaction of the software developers. The overall well being and job satisfaction of the software developers were interrelated. These findings were quite well in line with those obtained from literature review findings. Introduction Successful execution of software development processing is integral for the entire human race. Software development, an approximately sixty four year old process now, is an ever expanding field which is undergoing constant change and adoption. Customer satisfaction and enhanced usability are two of the main goals that software development environments focus upon. The report is aimed at highlighting the relationship among the agility, well being and job satisfaction of software developers in a software development arena. The current research, specifically, revolves around the relationships that may exist between agility, job satisfaction and wellbeing with respect to software developers. In order to research the stated factors a set of software developers belonging to different development environments were contacted and inquired about their point of view. Their responses were assessed by making them respond to a pre-prepared set of questions. The analysis done on the data collected via questionnaire responses was safeguarded for future use. In the meantime rigorous literature review was conducted in order to assess the past researches and the level of correlation obtained by past researchers among the variable factors under discussion here. An analysis of the findings both via literature review and analytical approach is then presented henceforth. Background The background of this research revolves around the recent and rising aptitude of software development and certain factors relative to it. The infusion of technology into human lives has reached unreachable limits and there is absolutely no going back from this stage. These advancements are all at the disposal of the software that make the techies work. The making of all kinds of software is the key factor in the running of any technological device. Making of the software may be affected by a number of factors. These may include the personalities of the software developers working on their making, the circumstances within the arena, the available resources, the methodologies being adopted to create the software and the job satisfaction of the software developers etc. In order to optimize the successful conduction of software development it is inevitable to find out the affects of certain factors upon the core development processes. Data Acquisition and Description The method of data collection deployed for this research was through questionnaire. The questionnaire was designed to explore certain traits and characteristic related to software developers’ profession and their practices. The questionnaire was named as “2013 Software Developer Census”. The population addressed through this questionnaire was of diversified national background. The entire questionnaire comprised of more than two hundred questions that are bunched up in about 91 categories. The sample space included 98 candidates (software developers) who responded with a filled questionnaire. As it was already mentioned in the questionnaire that there was no obligation of attempting the questions in full, therefore there was a considerable number of missing responses in certain questions. The data file is formulated as SPSS data file (.sav). The software package used for data analysis was SPSS. In line with the purpose of this research, the report aims to study only few of the variables from the whole data file. The categories under study are, Job Satisfaction: This category is composed of 3 questions with Likert Scale responses given scores from 1 – 5, for 1 as strongly disagree and 5 as strongly agree. The “Job Satisfaction Index” is an aggregate of scores of responses. The value of “Job Satisfaction Index” hovers between the ranges of 3 – 15 with an expected mean value of 9. Psych Wellbeing: This category is composed of 5 questions with 7 Likert Scale responses. The scores fall among 1 – 7 (where 1 represents strongly disagree, 4 represents neutral and 7 represents strongly agree). The “Psych Wellbeing Index” represents sum of scores of responses that is ranged in 5 – 35 with an expected mean value of 20. Agility: The category includes 9 questions with 5 Likert Scale responses each. The “Agility Index” is a calculated variable that represents the extent of agility of a software developer in a quantitative manner. The acquired scores are added to calculate the value of “Agility Index”. With 9 questions the minimum and maximum expected scores were 9 and 45 respectively. The expected mean value is 27. The three variables mentioned above are further recoded and processed as different variables. The names of recoded variables are JSCoded (for Job Satisfaction Index), WBCoded (for Wellbeing Index) and AICoded (for Agility Index). These variables are re-coded on the basis of binary categorization of disagreement and agreement with respective values of 1 and 2. The expected mean value for each original variable is used to create the classes. This is done to achieve groups in data for testing the extent of significance regarding disagreement or agreement of the sample under study. Considering the different magnitudes of all three variables (“Job Satisfaction Index”, “Psych Wellbeing Index” and “Agility Index”) an attempt was made to stabilize the magnitude of scores by converting them into percentage weights. This was proved to be effective in comparing the mean values of all three variables and to conclude respective results. Research Questions The purpose of this research was to explore the extent of three basic factors regarding the professional environment and factors that may affect software developers. The research encompasses a sample extracted from international population. Keeping the aim of this study intact following are few research questions that had been formulated after the initial study of the scenario. 1. Whether the software developers are generally satisfied with their current jobs? If yes or no, then to what extent? How would it be proved? The implications on other factors? 2. Similarly In what state of psychological wellbeing the software developers generally are? How would it be rated? Positively or Negatively? What are its implications 3. What is the general level of agility that is exhibited by software developers currently? Is it good or bad? What are its effects on other factors? 4. Are there any significantly affecting relationships among the above mentioned factors? Any specific trends of interaction? 5. Whether there may be other external (other than these three) factors that may affect the results of this study? The hypotheses are formulated and tested accordingly under the light of these research questions supported with the elementary analyses of descriptive statistics and profile analyses of related graphs. Literature Review Software development may be affected by a number of factors. These factors may be related to the specifications of the software being developed itself, to the modules upon which the maturing of the software is being targeted and to several conditions with respect to the software developers themselves. Job satisfaction of the software developers, their well being and agility may be a few of among the hundreds of factors that may have affects on the software development environment and thus o the software development process itself. The researches done in the recent past with respect to the above mentioned factors and their correlation with the process of making software is presented forth as follows: Agile development processes and job satisfaction among software developers in the software development environment have often been mentioned to have a direct relationship. Tessem and Maurer (2007, p.56) mention, "Job satisfaction and motivation are claimed to be one of the main effects of using agile software development methods, and this is confirmed by Melnik and Maurer in a comparison of agile and non-agile software developers." (Tessem, Maurer, 2007, p.56) There have been a number of studies that emphasize that the implementation of agility in the software development environment directly affects the software developers eventually resulting in a constructive environment where the people involved get highly motivated, satisfied and content with the situations at their work. (Tessem, Maurer, 2007, p. 56) Gallivan (2003, p. 447) mentions the relationship between the well being of a software developer and his job satisfaction. The freedom and the scope to exercise creativity, when availed by a software developer, usually result in the software developers job satisfaction, "Software developers who are innovators will report higher levels of job satisfaction following mandatory adoption of client/server development, compared to adaptors." (Gallivan, 2003, p. 447) One of the interesting findings regarding software development from the existing research has been that it is the job category that is the most influenced by the job satisfaction, successful agile implementation and the well being of the individual worker (software developer in this case) among all the job categories ever present. (Linberg, 1999, p.178) Overall dissatisfaction of the software developer thus hampered well being of the constructors of software leads to immense adverse effects on job satisfaction. (Parayitam et al., 2010, p. 347) Relationship between agile software development and importance of job satisfaction (Dybe, Dingsoyr, 2012, p.849) has been inconclusive in certain study domains as quoted below: "The effect on work practices and job satisfaction of using agile and traditional methods has not been established conclusively. Some studies have found that work practice is more standardized when agile methods are used and that job satisfaction is greater." (Dybe, Dingsoyr, 2012, p. 850) Well being of software developers comes automatically with high achievement motivation. "The degree of job satisfaction of the workers with high achievement motivation exceeds that of workers with low achievement motivation." This naturally implies that well being and job satisfaction of software developers are directly related to each other. (Chen, 2008, p.107) Dingsoyr et al. in ‘A decade of Agile Methodologies’ (2012, p.1214) mention software development agility as the ability of software developers to respond to user requirements during the project life cycle. This ability is thus directly related to the software developer’s well being and job satisfaction. Modern traditions in software development and agile practice (pedrycz et al., 2011, p.741) define the relationship of job satisfaction with them as being unpredictable. Methods and Finding/Themes Statistical methods that are used to perform data analysis include, Elementary analysis using descriptive statistics. Profile Analysis using related graphs. Hypotheses formulation. T-tests for independent and paired samples. Chi-squared test of independence. Correlation. The data is analyzed using SPSS and the tables/figures are included in the appendices. The deployment of above mentioned statistical methods revealed a number of findings. These findings are basically based on the analysis of three key variables that are related to software developers’ job satisfaction, psychological wellbeing and exhibited extent of agility. The text below highlights the findings with proper references of respective tables and figures. The analysis of descriptive statistics (See Table 1.0, Appendix A) primarily discloses a few trends regarding the three variables in focus. The Job Satisfaction Index seems to have a normal distribution with coinciding mean, median and mode values that are 10 approx. As compared to the expected mean value i.e. 9, the current mean seems to be inclined more towards the notion agreement regarding job satisfaction. The histogram (See Figure 1.0, Appendix B) asserts the assumption of normal distribution. The Psychological Wellbeing Index does not seem to have an ideal normal distribution (See Figure 1.1, Appendix B). The mean and median are almost coinciding i.e. 23 approx. The mode (22) < median (23) < mean (23.19). However these values are higher than the expected mean value i.e. 20, highlighting the fact that generally the software developers exhibited a notion of agreement regarding their wellbeing. The higher standard deviation indicates less data consistency in this variable as compared to the other variables under study. The fluctuations in the responses may be the reason behind this dispersion. It seems that overall the responses regarding well being are quite shattered. The Agility Index contains the relationship of (See Table 1.0, Appendix A) mean (29.6) < median (30) < mode (32). The distribution appears to be almost normal (See Figure 1.2, Appendix B). The mean value is greater than the expected mean i.e. 27. Therefore apparently it can be said that the extent of agility as per the responses is towards positive side i.e. agreement. The graphical comparison of mean values of all three variables can be seen through box plot (See Figure 1.3, Appendix B). The different magnitude of these variables is synchronized by obtaining the percentage weights of responses in three different variables namely JSIndex, WBIndex and AIndex. The initial findings stated above seem to be emphasized over through the box plot under study. An effect of higher data inconsistency can be seen in WBIndex that represents Psychological Wellbeing Index. The mean values of Job Satisfaction Index and Agility are seen higher as compared to the Wellbeing Index. The Paired sample t-test (See Table 1.9, Appendix A) is performed using all three pairs to investigate the trend explained above. The variables with adjusted magnitude are used for the purpose. Let µ1 and µ2 represent the means of first and second variable in the pair. The null and alternative hypotheses formulated for all the pairs is given as, H0: µ1 = µ2 H1: µ1 ≠ µ2 It is revealed through the study of Table 1.9 (Appendix A) that null hypothesis is rejected for pairs Job Satisfaction and Wellbeing and Agility Index and Well being with p-value < 0.05 showing significant difference in mean values. While the null hypothesis is not rejected for the pair Job Satisfaction and Agility Index due to non significant difference in mean values (See Table 1.9, Appendix A). These findings support the trend observed through box plot ((See Figure 1.3, Appendix B). In an attempt to answer the research question formulated above the sequential discussion is presented in the text below. In order to prove the apparent inclination of respondents towards agreement regarding Job Satisfaction, Well being and Agility the hypothesis is formulated by introducing binary groups/classification in data. The grouping technique is already discussed in the text above. The extent of agreement against the extent of disagreement is tested through independent sample t-test using following hypothesis with level of confidence = 95%. Let µ1 and µ2 represent the mean values of the responses towards disagreement and agreement. The null and alternative hypotheses as per the research questions formulated collectively for all three variables would be as follows, H0: µ1 = µ2 (There is no difference between mean values of both the groups) H1: µ1 ≠ µ2 (The mean values of both the groups are significantly different) The study of Tables 1.11, 1.13 and 1.15 (See Appendix A) reveal the results that with 95% level of significance and p-value (0.000) < α (0.05) the null hypothesis is rejected and it is concluded that there is a significant difference between the mean values of both the groups. The positive t-values and Confidence Intervals indicate that µ1 < µ2 in all cases. This again supports the initial finding that the extent of agreement is higher than the extent of disagreement. In order to analyze the interdependence of the three variables respective correlations are studied through Pearson Correlation Coefficient. The correlation (r=0.47) between Wellbeing Index and Job Satisfaction Index is found to be significant with p-value (0.00) < 0.05 (See Table 1.16, Appendix A). This result is supported by the analysis of correlation (r=0.62) between the adjusted variables JSIndex and WBIndex which is found to be significant as well (See Table 1.16, Appendix A). The other two pairs do not seem to have any significant correlation. The initial findings related to interdependence of the three factors are tested through Chi-Squared test of Independence (See Tables 1.2, 1.4 and 1.6, Appendix A) for the test results of all the pairs. The variables used are the coded versions of original variables in order to introduce two groups of two extents of agreement and disagreement. In order to conduct analysis in line with the answers of the formulated research questions the null and alternate hypotheses for testing the dependence of all pairs can be given as, H0: The pair factors are independent. H1: The pair factors are dependent. According to both the approaches i.e. Critical Value Approach and ­p-value approach the test results are narrated as follows, For Job Satisfaction and Psychological Well being: With chi square statistic χ2 = 12.65 > critical value (3.84), level of significance (α = 0.05), p-value (0.00) < α and degrees of freedom (df =1) the null hypothesis is dependent is rejected (See Table 1.2, Appendix A). For other pairs the null hypothesis is not rejected showing no dependency (See Tables 1.4 and 1.6). Discussion and Implication The findings of the statistical analysis were as follows: Most of the people are satisfied with their well being, job and the agility index of most of them is positive. Job satisfaction and well being have a significant relationship. Whereas agility and job satisfaction do not have a direct relationship. Also, Agility and well being also do not have direct relationship. Responses to well being are not consistent. The responses obtained are highly varied. Therefore the aspect of well-being may still be kept under observation. The mean values magnitudes of Job satisfaction and Agility are higher than the mean value of magnitude of well-being. This supports the fact that regarding job satisfaction and agility the responses of people are clearly inclined towards agreement whereas this is not the case in well-being. The research of literature review showed that According to Tessem and Maurer (2007, p.56) regarding job satisfaction that it is being affected by agility is contrary to the statistical findings done in this research which showed there is no direct relationship between two. However, Dybe, Dingsoyr (2012, p.849) stated that agility has no direct effect upon job satisfaction and this finding supported the results obtained in the analytical results obtained via this report. Pedrycz et al. (2011, p. 741) seconded the opinion that the relationship between agile development and job satisfaction is unpredictable. Gullivan (2003, p.447) mentioned that job satisfaction and ideal innovative working environments have a positive effect upon each other. This research finding is synchronized with the results obtained by statistical analysis. This factor has been curtailed in the supplementary questions of well being category. The results are matching. Parayitam et al. (2010, p.347) also seconds the same effect. Linberg (1999, p.178) partially agreed to the statements discussed above. Chen (2008, p.107) narrated a highly correlation between well being and job satisfaction. Conclusion Conclusions aligned with research questions may be as follows: 1. Software developers seem generally satisfied with their current jobs with a fairly positive extent. 2. Job satisfaction and well being are proved to be interrelated and effecting each other. 3. With respect to psychological development the software developers gave an overall positive output but their response was a bit shaky due to high data dispersion. 4. Wellbeings relationship with job satisfaction is significant. 5. Positive trends are exhibited by the respondents regarding extent of agility. There seems to be no direct correlation of agility with the other two factors. 6. These three factors which have been focused upon may not be the only factors that would provide true and realistic results. The addition of more factors would improvise the authenticity of the results even more. These factors may be: the job satisfaction index exhibited by gender (See Table 1.18) References Chen, L. H. 2008. Job satisfaction among information system (IS) personnel.Computers in Human Behavior, 24(1), 105-118. Dybå, T., & Dingsøyr, T. 2008. Empirical studies of agile software development: A systematic review. Information and software technology, 50(9), 833-859. Dingsøyr, T., Nerur, S., Balijepally, V., & Moe, N. B. 2012. A decade of agile methodologies: Towards explaining agile software development. Journal of Systems and Software. Damiani, G. C. E., & Succi, M. S. G. 2007. Agile Processes in Software Engineering and Extreme Programming. Gallivan, M. J. 2003. The influence of software developers’ creative style on their attitudes to and assimilation of a software process innovation. Information & Management, 40(5), 443-465. Kurt R, L n.d., Software developer perceptions about software project failure: a case study, The Journal Of Systems & Software, 49, pp. 177-192, ScienceDirect, EBSCOhost, viewed 27 January 2013. Linberg, K. R. 1999. Software developer perceptions about software project failure: a case study. Journal of Systems and Software, 49(2), 177-192. Parayitam, S., Desai, K. J., Desai, M. S., & Eason, M. K. 2010. Computer attitude as a moderator in the relationship between computer anxiety, satisfaction, and stress. Computers in Human Behavior, 26(3), 345-352. Pedrycz, W., Russo, B., & Succi, G. 2011. A model of job satisfaction for collaborative development processes. Journal of Systems and Software, 84(5), 739-752. Silvia T., A, Marta, G, & Natalia, J n.d., How do personality, team processes and task characteristics relate to job satisfaction and software quality?, Information And Software Technology, 51, pp. 627-639, ScienceDirect, EBSCOhost, viewed 27 January 2013. Tessem, B., & Maurer, F. 2007. Job satisfaction and motivation in a large agile team. Agile Processes in Software Engineering and Extreme Programming, 54-61. Viraj, S, & Chandana, G n.d., Employee perception towards electronic monitoring at work place and its impact on job satisfaction of software professionals in Sri Lanka, Telematics And Informatics, 29, pp. 233-244, ScienceDirect, EBSCOhost, viewed 28 January 2013. Appendix A Descriptive Statistics Table 1.0. Descriptive Statistics Job Satisfaction Wellbeing Agility N Valid 92 89 89 Missing 6 9 9 Mean 10.05 23.19 29.60 Median 10.00 23.00 30.00 Mode 10 22 32 Std. Deviation 1.529 6.237 4.499 Range 9 30 23 Minimum 5 5 17 Maximum 14 35 40 Test Results Table 1.1. JSCoded * WBCoded1 Crosstabulation Count WBCoded Total 1.00 2.00 JSCoded 1.00 15 12 27 2.00 11 50 61 Total 26 62 88 Table 1.2. Chi-Square Tests of Independence regarding Job Satisfaction and Wellbeing Value df Asymp. Sig. (2-sided) Pearson Chi-Square 12.659a 1 .000 Continuity Correctionb 10.921 1 .001 Likelihood Ratio 12.159 1 .000 N of Valid Cases 88 a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 7.98. b. Computed only for a 2x2 table Table 1.3. JSCoded * AICoded Crosstabulation Count AICoded Total 1.00 2.00 JSCoded 1.00 6 20 26 2.00 12 50 62 Total 18 70 88 Table 1.4. Chi-Square Tests of Independence regarding Job Satisfaction and Agility Value df Asymp. Sig. (2-sided) Pearson Chi-Square .156a 1 .693 Continuity Correctionb .011 1 .916 Likelihood Ratio .153 1 .695 Fishers Exact Test Linear-by-Linear Association .154 1 .695 N of Valid Cases 88 a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 5.32. b. Computed only for a 2x2 table Table 1.5. WBCoded * AICoded Crosstabulation Count AICoded Total 1.00 2.00 WBCoded 1.00 4 21 25 2.00 14 47 61 Total 18 68 86 Table 1.6. Chi-Square Tests of Independence regarding Wellbeing and Agility Value df Asymp. Sig. (2-sided) Pearson Chi-Square .518a 1 .472 Continuity Correctionb .183 1 .669 Likelihood Ratio .539 1 .463 Fishers Exact Test Linear-by-Linear Association .512 1 .474 N of Valid Cases 86 a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 5.23. b. Computed only for a 2x2 table T-Test Table 1.7. Paired Samples Statistics Mean N Std. Deviation Std. Error Mean Pair 1 JSIndex .6738 98 .20351 .02056 WBIndex .6015 98 .25648 .02591 Pair 2 JSIndex .6738 98 .20351 .02056 AIndex .6736 98 .24034 .02428 Pair 3 WBIndex .6015 98 .25648 .02591 AIndex .6736 98 .24034 .02428 Table 1.8. Paired Samples Correlations N Correlation Sig. Pair 1 JSIndex & WBIndex 98 .626 .000 Pair 2 JSIndex & AIndex 98 .544 .000 Pair 3 WBIndex & AIndex 98 .435 .000 Table 1.9. Paired Samples Test Paired Differences t df Sig. (2-tailed) Mean Std. Deviation Std. Error Mean 95% Confidence Interval of the Difference Lower Upper Pair 1 JSIndex - WBIndex .07224 .20449 .02066 .03125 .11324 3.497 97 .001 Pair 2 JSIndex - AIndex .00020 .21439 .02166 -.04278 .04319 .009 97 .993 Pair 3 WBIndex - AIndex -.07204 .26450 .02672 -.12507 -.01901 -2.696 97 .008 Independent Samples Tests Table 1.10. Group Statistics Job Satisfaction JSCoded N Mean Std. Deviation Std. Error Mean Job Satisfaction 1.00 28 8.29 1.049 .198 2.00 64 10.83 .952 .119 Table 1.11. Independent Samples Test Levenes Test for Equality of Variances t-test for Equality of Means F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Job Satisfaction Equal variances assumed .427 .515 -11.424 90 .000 -2.542 .223 -2.985 -2.100 Equal variances not assumed -10.995 47.332 .000 -2.542 .231 -3.008 -2.077 Table 1.12. Group Statistics Wellbeing WBCoded N Mean Std. Deviation Std. Error Mean Wellbeing 1.00 27 15.81 4.386 .844 2.00 62 26.40 3.655 .464 Table 1.13. Independent Samples Test Levenes Test for Equality of Variances t-test for Equality of Means F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Wellbeing Equal variances assumed .425 .516 -11.812 87 .000 -10.588 .896 -12.370 -8.807 Equal variances not assumed -10.993 42.452 .000 -10.588 .963 -12.532 -8.645 Table 1.14. Group Statistics AICoded N Mean Std. Deviation Std. Error Mean Agility 1.00 18 23.00 1.940 .457 2.00 71 31.27 3.242 .385 Table 1.15. Independent Samples Test Levenes Test for Equality of Variances t-test for Equality of Means F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Agility Equal variances assumed 8.237 .005 -10.332 87 .000 -8.268 .800 -9.858 -6.677 Equal variances not assumed -13.833 44.210 .000 -8.268 .598 -9.472 -7.063 Correlations Table 1.16. Correlations Job Satisfaction Wellbeing Job Satisfaction Pearson Correlation 1 .473** Sig. (2-tailed) .000 N 92 88 JSIndex WBIndex JSIndex Pearson Correlation 1 .626** Sig. (2-tailed) .000 N 98 98 **. Correlation is significant at the 0.01 level (2-tailed). Table 1.17. Group Statistics What is your gender? N Mean Std. Deviation Std. Error Mean Job Satisfaction Female 18 9.39 1.461 .344 Male 67 10.25 1.531 .187 Table 1.18. Independent Samples Test Levenes Test for Equality of Variances t-test for Equality of Means F Sig. t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Job Satisfaction Equal variances assumed .003 .955 -2.148 83 .035 -.865 .403 -1.666 -.064 Equal variances not assumed -2.207 27.887 .036 -.865 .392 -1.668 -.062 Appendix B Figure 1.0. Histogram (Job Satisfaction Index) Figure 1.1. Histogram (Wellbeing Index) Figure 1.2. Histogram with Normal Curve (Agility Index) Figure 1.3. Box Plot representing comparison of mean values Read More